ASCI course A1 :
Advanced Pattern Recognition

Pattern Recognition Laboratory
Delft University of Technology, Delft

This graduate course is intended for Ph.D. students that face [research] problems that may partly be solved by automatic classifiers. Statistical pattern recognition and machine learning techniques are examined with an emphasis on the generalization capabilities of the learning systems.

The course is given primarily for Ph.D. students of the Advanced School for Computing and Imaging [ASCI]. Postdoctoral researchers and external Ph.D. students may also attend against a moderate fee. It is assumed that the student has successfully finished introductory courses on linear algebra, probability theory, and statistics, but not specifically in pattern recognition.

The aim is to give the course once a year in a one-week block. More than half of the time will be used for hands-on experience using Matlab [and in particular PRTools]. The students are advised to work in groups of two on these Matlab exercises for which they should arrange a laptop with the necessary software etc. themselves. The course will be concluded by a small project together with a written report to be handed in after the course.

“Advanced” in What Sense?

The course is deemed “advanced” because of the sheer volume of techniques, methods, and concepts that the course ASCI covers. Though we will present both old and new, we cannot [and don’t want to] cover all the current hypes and novelties in depth, especially if they require, for instance, dedicated implementations. [So we would typically touch upon neural networks, but not go in depth when it comes to deep learning.]

A final word for the people that happened to follow the master’s course on pattern recognition at TU Delft [IN4085] or basically any other course in machine learning or pattern recognition: yes, there is probably a large overlap, but unless you are really unreceptive, you will get new insights and/or deeper understandings out of the APR course.


ASCI course A1 : Advanced Pattern Recognition